DSpace 7 community and collection statistics are inaccurate

[alanorth]

Describe the bug
Statistics displayed on the "Statistics" page of a community or collection are inaccurate (in my opinion). They seem to only reflect views of the community or collection page itself, rather than counting child communities, collections, items, and bitstreams.

To Reproduce
Steps to reproduce the behavior:

1. View the "Statistics" page for a community or collection, take note of the UUID
2. Use Solr to perform a search for statistics records directly belonging to that UUID, for example: id:d09e6100-df60-4280-9ff3-f3fea5ea4e6b. These are views of the community/collection page itself.
3. Use Solr to perform a search for statistics records belonging to children of that UUID, for example: owningComm:d09e6100-df60-4280-9ff3-f3fea5ea4e6b (or owningColl for a collection).

Expected behavior
The number of "views" of a community or collection should take into account all children of the community or collection, in addition to the community page itself. We already have the owningComm and owningColl fields in Solr managed by DSpace. We can use those to show more accurate statistics than we are currently doing, without much work.

For example, our DSpace repository has years of Solr statistics and currently shows this for the simple id query of a large community:

$ curl 'http://localhost:8983/solr/statistics/select?q=id:d09e6100-df60-4280-9ff3-f3fea5ea4e6b&rows=0'
{
...
  "response":{"numFound":2543,"start":0,"numFoundExact":true,"docs":[]
}}

But really there are many, many more if we include the child communities, collections, item pages, and bitstreams via the owningComm:

$ curl 'http://localhost:8983/solr/statistics/select?q=owningComm:d09e6100-df60-4280-9ff3-f3fea5ea4e6b&rows=0'
{
...
  "response":{"numFound":6461926,"start":0,"numFoundExact":true,"docs":[]
}}

In addition to this, at our institution we disaggregate statistics into "views" and "downloads" using the type field (see DSpace Constants¹):

• A "view" is something that is an HTML landing page for a community, collection, or item (types 2, 3, 4)
• A "download" is something that is a "view" of a bitstream (type 0) where the bundleName:ORIGINAL

So in Solr I can see this community had:

• 1872043 "views"
• 2582247 "downloads"

You can see how we count views and downloads in our standalone dspace-statistics-api indexer.

Related work

• DS-9142: prevent UUID exception by adding check for -1 value for the id field #9261

---

¹ See dspace-api/src/main/java/org/dspace/core/Constants.java

[alanorth]

This is a duplicate of #8572.